Adding parts for a greater whole

From oemagazine May 2003
01 May 2003
By Hubert Kostal

Integration is an inevitable evolution of the photonics process. From a systems perspective, it allows the same modules to be used for numerous applications, thus increasing usage and accelerating payback. The optimized performance gained from repetitive use of a module decreases or eliminates development and training time and lowers operations and manufacturing costs; moreover, systems can be designed to reuse constituent components for multiple functions, which reduces material cost and manufacturing complexity.

Today, many non-specialized photonic circuits are constructed by stringing together discrete optical components. As a consequence, engineers must achieve optical performance objectives and control assembly cost while dealing with materials incompatibilities and complex designs. Integration of multiple optical functions into a single device lowers cost-per-function and simplifies assembly. In telecom, these characteristics are critical for access/enterprise optical networks and network-wide optical performance monitoring. In photonic circuits for projection displays, optical metrology, and sensors for chemical, biological, and environmental applications, integration simplifies assembly, reduces overall cost, and increases packing density.

To be viable, an integrating technology must support a broad range of functional capabilities (i.e., provide multiple modular "building blocks"), allow those building blocks to be easily combined to accomplish more complex functions, and be suitable for volume manufacturing. Optical integration approaches featuring planar light-guide circuits (PLCs), micro-electro-mechanical systems (MEMS), and indium phosphide (InP) devices offer improvements over the discrete approach, but none meet all three requirements. To accomplish that requires changing the rules of the game.

Nano-optics does just that by accessing a physical realm of optical behavior not accessible by bulk optic components. Combining wafer-scale manufacturing, an optical building block fabric that can readily be reconfigured to provide a range of optical functions, and a capability for integration that is compatible with both, nano-optics meets all three requirements for an integrating technology for photonic integrated circuits (PICs).

into the nano-regime

When light passes through a grating or structure with critical dimensions much smaller than the wavelength of light, the physical interactions are significantly different than for bulk optics. Such gratings or structures are called nano-optics, ranging from tens to hundreds of nanometers in size, often specified to an accuracy of 10 nm or less. Much of the novel behavior of the devices comes from the control that materials and structure provide over the light beam.


Figure 1. A nano-optic chip consists of multiple optical layers surrounding a subwavelength grating structure.

Consider a nano-optical device for telecom (see figure 1). Such a device would consist of a silicon-dioxide layer less than 1-µm thick, a substrate of optical quality glass from 0.2-mm to 1.0-mm thick, and a grating structure with a period of 1 to 300 nm. Since the performance and functionality of such devices are based on both the shape of the grating structure and on the materials used, the technology can easily be applied to various wavelength ranges, from the UV to IR spectral regions.

Nano-optic behavior can be described by rigorous general optical physics, and, in the extreme, quantum and single-electron effects. As an example, the diffraction for a nano-optic grating is very different than that for classical grating structures because the nano-grating period is shorter than the wavelength of the incident light. The conventional grating equation for light impacting a surface at normal incidence is:

a sinθm=mλ

where a is the period of the grating, m is the diffraction order, θm is the diffraction angle, and λ is the wavelength of the light. When a is less than λ, nearly all optical energy is diffracted into the zeroth order, which implies uniform and efficient behavior across a broad range of wavelengths. Practically speaking, this means that a nano-optical polarization beamsplitter/combiner or polarizer would provide relatively flat optical performance across a range of incident angles. Such behavior has benefits both for application architectures and automated assembly operations.

Nano-optics can be designed to manage polarization, phase, wavelength, and other optical properties by appropriate selection of materials and structural parameters. Examples of components include polarizers, polarization beam splitters and combiners, waveplates, wavelength filters, photodetectors, lasers, lenses, and various antireflective or all-reflective coatings. Because the structural parameters can be continuously varied, nano-optics can be designed to be application specific; for instance, the target percentage power reflected from a nano-optic filter can be adjusted from full reflection to only partial reflection.

There are a number of different ways to construct these structures, ranging from molecular self-assembly to e-beam lithography. These methods work well for research applications in which yield and volume are not issues, but in commercial applications, flexibility, repeatability, volume, and low cost are key. Nano-pattern transfer manufacturing methods offer a high-speed, low-energy alternative by using molds to pattern a prepared wafer that is subsequently etched using standard semiconductor processing techniques. With a change of molds to specify different structures, the same manufacturing line can be used to create a full range of nano-optic devices.


Figure 2. A 3-D nano-optical structure forms the basis for a PIC.

Because the optical effects created by nano-optics depend on structures that are fabricated using semiconductor manufacturing techniques, these functions can be readily combined—both by varying optical functionality spatially on a single wafer level and by processing the same wafer multiple times to create multiple layers of nano-optic structures. This allows the creation of complex, 3-D structures (see figure 2).

thinking about integration

A PIC integrates two or more discrete optical functions in a single device at chip level. A simple PIC combines two or more serial functions, reducing the component count in an optical circuit. In more sophisticated applications, a single PIC will encompass all of the optical processing functions in the optical circuit—for example as an optical performance modulator or optical add/drop multiplexer.

Integration can be classed as monolithic and hybrid. In monolithic integration, a single technology is combined with itself at a chip level to realize complex optical functions. In hybrid integration, the core technology is combined with other technologies during fabrication—as opposed to during later assembly—to achieve higher-level functions.


Figure 3. A nano-optic waveplate combined with a polarization beamsplitter can measure differential polarization and phase response in a compact polarimeter.

Using nano-pattern transfer manufacturing, we can achieve monolithic integration by layering one structure on top of another. Because the optical functionality occurs in the very thin nano-optic layer, spacing between layers does not require the same optical length as it does for standard discrete optics. More important, variations in nano-structures can be created within a single layer. These variations can be continuous to take advantage of the statistical nature of the interaction of light and nano-optics structures, or they can be pixilated to allow multiple parallel beams of light to be differentially processed. Combinations of layering and pixilation allow complex differential processing of multiple parallel beams of light. In addition, light can be spatially redirected within an integrated nano-optic device via a waveguide layer (see figure 3).

Hybrid integration of nano-optics and other materials increases the richness of functionality available. Because structures are fabricated using a wafer-based process, these manufacturing steps can be easily integrated with other manufacturing operations to take advantage of optical properties that can be created by non-nano-optic means. For example, nano-optic structures can be fabricated on arbitrary substrates so long as materials are matched to account for relevant thermal expansion. Creating a nano-optic polarizer on a garnet substrate, for instance, would yield a compact semi-isolator. Combining dynamic layers such as liquid crystal, MEMS, or some other material whose optical properties can be electrically or magnetically controlled with a nano-optic structure can produce tunable functions and switches.

Nano-optics as a technology is compatible with the general requirements for a core fabric for PICs, combining a broad set of optical building blocks with a manufacturing process amenable to their combination. In addition to offering design flexibility, the technology is cost competitive, it is compatible with other approaches such as PLC, MEMS, and InP, as well as with wafer-based manufacturing processes. By making available a wide range of unique optical properties and arrangements, nano-optics opens the possibility of new component designs and architectures. oe


Hubert Kostal
Hubert Kostal is vice president at NanoOpto Corp., Somerset, NJ.

Recent News
PREMIUM CONTENT
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research